The sensory and perceptual processes whereby the acoustic waveform of speech is transformed into a sequence of phonetic units are to be studied. New methods for speech measurement that will be more accurate and/or more automatic than current methods will be developed and/or evaluated. A set of recordings representing difficult problems for current speech analysis methods will be established and carefully analyzed using hand-intervention. The adequacy of proposed new methods will be empirically validated using these results. Issues related to spectral estimation, pitch extraction, formant labeling and tracking, measures of loudness, goodness, and durations of speech sounds will all be addressed. The factors determining the salience of the acoustic signal and its segmentation into perceptually relevant units will be investigated. Data will be collected through speech measurements and listening experiments with natural and synthetic speech to test the concept of perceptual target zones and determine their exact structure. A mathematically defined sensory-perceptual transformation will be further developed to include loudness and goodness and other factors of importance. Extensive empirical work will be directed to the acoustic correlates of articulatory parameters and to the effects of speaking rate and stress on the acoustic characteristics of speech sounds. Finally, speech perception by the hearing- impaired will be modeled so that a more precise description of their difficulties can be achieved. A large amount of new data related to speech perception will be accumulated. Computer graphics will allow verification of the data and visualization of the concepts of the auditory-perceptual theory of phonetic recognition that provides a quantitative framework for the research. An improved understanding of speech perception and its disorders is expected with resultant implications for remedial treatments.